067 Configure 5ca2e2f6

Embedding Configuration

Configure embedding models for vector search.

Volcengine Doubao (Recommended)

{
  "embedding": {
    "dense": {
      "provider": "volcengine",
      "api_key": "your-volcengine-api-key",
      "model": "doubao-embedding-vision-250615",
      "dimension": 1024,
      "input": "multimodal"
    }
  }
}

Parameters

Available Models

Getting Volcengine API Key

1. Visit Volcengine Console
2. Navigate to Ark service
3. Create an API key
4. Copy the key to your configuration

Environment Variable

export VOLCENGINE_API_KEY="your-api-key"

Then in config:

{
  "embedding": {
    "dense": {
      "provider": "volcengine",
      "model": "doubao-embedding-vision-250615",
      "dimension": 1024
    }
  }
}

Programmatic Configuration

from openviking.utils.config import EmbeddingConfig, DenseEmbeddingConfig

embedding_config = EmbeddingConfig(
    dense=DenseEmbeddingConfig(
        provider="volcengine",
        api_key="your-api-key",
        model="doubao-embedding-vision-250615",
        dimension=1024,
        input="multimodal"
    )
)

Multimodal Support

With input: "multimodal", OpenViking can embed:

• Text content
• Images (PNG, JPG, etc.)
• Mixed text and images

# Multimodal embedding is used automatically
await client.add_resource("image.png")  # Image embedded
await client.add_resource("doc.pdf")    # Text + images embedded

Troubleshooting

API Key Error

Error: Invalid API key

Check your API key is correct and has embedding permissions.

Dimension Mismatch

Error: Vector dimension mismatch

Ensure the dimension in config matches the model's output dimension.

Rate Limiting

Error: Rate limit exceeded

Volcengine has rate limits. Consider:

• Batch processing with delays
• Upgrading your plan

Related Documentation

• Configuration - Main configuration
• LLM Configuration - LLM setup
• Resources - Adding resources